https://ph02.tci-thaijo.org/index.php/TJOR/issue/feed Thai Journal of Operations Research : TJOR 2021-06-29T09:45:17+07:00 Associate Prof Dr.Kannapha Amaruchkul orjournal.th@gmail.com Open Journal Systems วารสารไทยการวิจัยดำเนิงาน https://ph02.tci-thaijo.org/index.php/TJOR/article/view/241957 Fleet assignment problem with integer linear programming (A case study : Nok air airline) 2021-05-31T12:00:41+07:00 Ratee Bojaras ratee.b@ubu.ac.th Piyatida Saensao ratee.b@ubu.ac.th <p>In this research, we use integer linear programming to model the fleet assignment problem. The purpose of this model is to assign the most appropriate fleet types &nbsp;of&nbsp; domestic flight schedules in Nok Air airline and to find the minimized operating cost. In addition, we have determined the optimal number of aircraft grounded overnight at each airport. The results revealed that we found the optimal solution by LINGO with the minimized operating cost and the optimal number of aircraft grounded overnight at each airport.</p> 2021-05-31T11:51:29+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/242191 Determining the Optimal Base Stock Level when Demand and Supply Disruption Length are Discretely Distributed 2021-05-31T12:00:42+07:00 Chirakiat Saithong chirakiat@eng.src.ku.ac.th Natthawut Wichianpong natthawut.wi@ku.th <p>Abstract</p> <p>In the face of possibility of supply disruption, an organization needs an approach to deal with the problem. This research work uses an inventory holding approach to address the supply disruption problem under a periodic review base stock inventory system. The objective of this research work is to determine the optimal base stock level, which yields the minimum total costs per unit of time. Considering both demand and supply disruption length as discrete distributions can help an organization derives the optimal base stock level when both demand and supply disruption length do not fit with any prevailed distribution, which fills the gap in the literature. In the numerical experiment section, using a wide range of parameters, it is found that the optimal base stock level and the minimum expected total costs per unit of time can be determined.</p> <p>&nbsp;</p> <p><strong>Keywords</strong>: Supply disruption, Inventory holding, Base stock system, Periodic review inventory system.</p> 2021-05-31T11:51:51+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/242677 Measuring the Efficiency of Public Service Sector Banks in India Using Two-Stage Closed System DEA approach 2021-05-31T12:00:43+07:00 Ms. Jenifer Christinal jeniferchristinal24@gmail.com Dr. P. Mariappan mathmari@yahoo.com Dr. Dinesh Dave daveds@appstate.edu <p>This paper examines the efficiency of public service sector banks in India using two stage Data Envelopment Analysis [DEA] technique. The proposed model investigates the efficiency of banks of with various input and production standards in each level. While comparing banks, it was determined that some banks are efficient in their profit earning, whereas other banks are efficient in functioning smoothly. The methodology defines the profit efficiency in stage 1 and performance effectiveness in phase 2 of the selected public sector banks in India. The study demonstrates the comparative assessment and efficiency rankings among the selected bank in India.</p> 2021-05-31T11:52:55+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/243329 Topic Modeling and Text Classification for Xenophobia on Twitter during COVID-19 2021-06-01T14:02:54+07:00 Budsabong Kachintararojn budsabong.kac@stu.nida.ac.th Duanpen Teerawanviwat duanpen@as.nida.ac.th Pachitjanut Siripanich siripanich52@as.nida.ac.th <p>This research presents a topic modeling using the Latent Dirichlet Allocation (LDA) and two classification algorithms, which are Random Forest and Support Vector Machine for Xenophobia on Twitter during COVID-19 with three methods of Word Embedding - TF-IDF, Word2vec and GloVe. The dataset contains Xenophobia on Twitter During COVID-19 around 1,000,000 tweets from Kaggle. This research aimed to build the topic model for Xenophobia on Twitter during COVID-19 and study suitable classification algorithms for Xenophobic or Non-Xenophobic and also provides a way for Twitter to filter out tweets that are potentially violent. The results of topic modeling was showed 3 unique important topics that given the lowest Perplexity of 114186.86 which are (1) Arrestment (2) Disgusting (3) Damnation. As the results of classification algorithms were showed that Random Forest when using TF-IDF given F1-Score, Recall and ROC were not different but Precision was the highest value of 0.35. Therefore, Random Forest when using TF-IDF was the most suitable algorithm for Xenophobic or Non-Xenophobic classify during COVID-19. In addition, Support Vector Machine when using Word2vec given the highest F1-Score and Precision value of 0.43 and 0.28, respectively. Therefore, Support Vector Machine when using Word2vec was the most suitable algorithm for Xenophobic or Non-Xenophobic classify during COVID-19.</p> 2021-05-31T11:53:20+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/243405 Proximal Policy Optimization on Casual Restaurant Raw Material Stock 2021-06-01T14:03:02+07:00 Nutthawat Ekthammanit ts.nutthawat@gmail.com Worapol Pongpech worapol@as.nida.ac.th <p>&nbsp; &nbsp;This research focuses on the Proximal Policy Optimization Algorithm of Reinforcement Learning to make a forecasting model of raw material stock in restaurants. Due to the restaurant's daily raw material stock ordered. There is a high deviation from the number of raw material stock used. The rest of the raw material stock become food waste. This causes fermentation and the formation of methane gas to rise to destroy ozone in the atmosphere. Which is the main cause of the greenhouse effect. This research investigated a One-Attribute Model and a Multi-Attribute Model. The dataset used in this research is synthetic data that use the normal distribution theory to make it. The model's performance was assessed using F-statistics, R-Square, and RMSE. We trained each model trained 12 million timesteps. The result showed that the Multi-Attribute Model would converge to the value optimization faster than the One-Attribute Model. We found that both models' accuracy is about 82 percent of the number of the test set where the number of the test set is 1,000. From this research, we can learn how to apply the Proximal Policy Optimization Algorithm of Reinforcement Learning make a forecasting model of raw material stock. To be able to forecast the number of raw material stock. As close as possible to the number of raw material stock used and it can reduce the number of food waste.</p> 2021-05-31T11:53:47+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/243909 Multicriteria Analysis for Method Selection in Old Coal Mine Pit Closure in East Kalimantan in Indonesia 2021-06-15T09:47:16+07:00 ฺBralee Hemearaswad bralee.h@ku.th Patcharaporn Yanpirat fengppy@ku.ac.th Sansanee Supapa fengsas@ku.ac.th <p>This research aims to propose the appropriate methodology for closing-down of 8 coal mine pits in East Kalimantan in complying with the mining laws of Indonesia. The Fuzzy Technique for Order of Preference by Similarity to Ideal Solution was employed to evaluate the decision criteria for selecting appropriate methods. In determining the importance weights of each criteria, Voting Analytic Hierarchy Process was used including to prioritize importance weights of the selected experts. The results revealed that 4 out of 8 coal mine pits were suitable for closure with material landfill from surrounding areas using normal operation methods. Two pits were suitable to close with material landfill from another potential mine pit and from the surrounding areas by an alternative mining dozer method, respectively. The rest of the two pits were suitable to close by utilizing the pits for other purposes.</p> 2021-06-15T09:47:16+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/243912 Cluster Analysis of Retirement Mutual Funds 2021-06-29T09:43:47+07:00 Sasiprapa Hiriote Hiriote_s@silpakorn.edu Chayanee Phunphiphat chayanee211140@gmail.com Pornpimon Kunphong kunphongim@gmail.com <p><span style="font-weight: 400;">A Retirement Mutual Fund (RMF) is a type of mutual fund that promotes savings and long-term investment for retirement planning. RMF offers a wide range of investment policies from low-risk fund that invests the most of its principal in government bonds to high-risk fund that focus its investment on stocks or gold. There are certain types of risk associated with RMF such as market risk, liquidity risk, concentration risk, and so on. In this work, we study the performance of some Retirement Mutual Funds offered by various financial institutions based on Sharpe ratio, Jensen’s alpha and Treynor ratio to identify groups of these RMFs using time series cluster analysis with Community Detection in Networks. The result provides some guidelines for investors to choose among these RMFs to minimize the investment risk by diversifying their portfolio.</span></p> 2021-06-15T09:49:01+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/243914 A Hybrid Firefly Algorithm to Minimize Total Cost of Earliness and Tardiness in Precast Production Scheduling 2021-06-29T09:45:17+07:00 Sivasit Wittayasilp sivasit.w@psu.ac.th Wanatchapong Kongkaew wanatchapong.k@psu.ac.th <p>This research studied the precast concrete production process in a stationary system. Each precast job was done through all production steps, but there may be interruptions in some production steps due to the available production time of the day. In this paper, a metaheuristic algorithm was developed to solve the precast production scheduling problem (PPSP) for a production sequence with a minimum total cost of earliness and tardiness penalties. This method combined the NEH heuristic of Nawaz, Enscore, and Ham, the firefly algorithm, and the new local search operations, called the hybrid firefly algorithm (HFA). To test the performance of the proposed HFA, the experiments were executed on 10 test problems and one real-world industrial case study with the runtime limits. The proposed HFA’s performance was also compared to five recent methods. The results showed that the HFA has a good optimizing ability for solving the PPSP and it reached the best-known solutions among all considered algorithms. Moreover, comparing to others, the proposed HFA significantly outperformed almost all comparison algorithms, except the HCS (hybrid cuckoo search). Thus, the proposed HFA became an effective and competitive metaheuristic method for the considered PPSP.<br>Keywords: firefly algorithm, scheduling, precast production, total cost of tardiness and earliness<br><br></p> 2021-06-15T09:50:40+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/243958 Applying a two-stage clustering-assignment model for an item-associated product family warehouse storage location problem 2021-06-15T11:18:02+07:00 Akkaranan Pongsathornwiwat akkaranan@as.nida.ac.th Kanitta Lawhawichitsak kanitta.dhl@gmail.com <p>The objective of this research was to re-design the storage location of raw materials and common products in the warehouse of a chain restaurant as the case study. An existing problem is unproductive travelling distance and non-value-added time-consuming order picking by the customer orders because of an inappropriate layout. As observed, the nature of the picking processes indicate that pickers pick the individual items by each brand along with common products that share together for all brands. To design new appropriate locations and product zones for each SKUs, this study proposed a Two-Stages Clustering-Assignment Model to deal with classifying product groups and determining product placements problems. Firstly, was to find Item-based associations of both product groups (items and all brands). Secondly, was then to determine placements for associated items by formulating the problem as warehouse storage location assignment problem so as to minimize picking distance. OpenSolver2.9.3_Beta_LinearWin was used to find the optimal solution. The result shows that the new configuration help reducing picking distance by 56% comparing to the current storage locations. This benefit proves more efficient as it reduces pickers from 11 to 8 regular staffs, which in turns saving labor costs about 39,000 baht per month or 468,000 baht per year.</p> 2021-06-15T09:52:18+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/243960 A Supervised Learning-Based Supplier Classification Model for Supplier Performance Evaluation Problem in SAP ERP 2021-06-15T12:31:17+07:00 Akkaranan Pongsathornwiwat akkaranan@as.nida.ac.th Narongsak Kowwilaisaeng narongsak.koww@gmail.com <p>The purpose of this research is to develop a supplier classification framework for supplier evaluation problem for overcoming the uncertain results from embedded linear scoring methods in SAP ERP. Datasets in this study are Purchase Order (PO) and Purchase Requisition (PR). PO data is used for modelling supplier classification models, and PR is used as an input data to evaluate the suppliers’ performances before making purchasing contracts. Three evaluation criteria, which are developed by interviewing with experts are quantity, quality, and delivery commitment. In addition, this research is developed the data extraction algorithm from SAP ERP system and data transformation in order for designing suitably datasets for supplier classification models. Nine classification models: Naïve bay, K-Nearest Neighbors, Support Vector Machine, Logistic Regression, Adaboost, Decision Tree, Random Forest, and Stacking SVM+DT and Stacking SVM+LR are applied to compare the performances to find the best classification model. In the analysis schemes, all material groups and each material group as suggested two levels of analysis are proposed in this study. Results show that Adaboost is the best classification model in both train and test datasets. Furthermore, Adaboost is also outperform others for both levels of analysis with accuracy performances for all material group (81.6%) and each material group (67.6), respectively.</p> 2021-06-15T09:54:12+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/243966 Determining Wood Sheet Cutting Pattern for Knockdown Furniture Using Heuristic Algorithms 2021-06-15T13:23:15+07:00 Punyanuch Kosutao punyanuch.ko@ku.th Chansiri Singhtaun chansiri.s@ku.ac.th Roongrat Pisuchpen fengros@ku.ac.th <p>The objective of this research is to determine the wood sheet cutting pattern for knock-down cabinets of a furniture factory in order to reduce wasted wood sheets.&nbsp; The number and size of wood planks which are the components of the cabinet are arranged on the wood sheets using an effective heuristic algorithm called best fit direct algorithm.&nbsp; The program for calculating and processing the arrangement of the wood planks is developed by Microsoft Excel VBA. This program receives the input data, which are number and size of wood planks. These wood planks are arranged according to the best fit direct algorithm. The number of wood sheets and wood chips are figured out and shown in Table form. The arrangement of wood planks on every wood sheet is shown in 2D picture that is convenient for applications. The numerical results show that using the developed program can reduce wood chips by 29.93 %.</p> 2021-06-15T09:55:18+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/243973 Optimization of Cash Logistics in A Cash Center 2021-06-15T09:57:21+07:00 jiraphat navakhunmongkol jiraphat.na@ku.th Juta Pichitlamken juta.p@ku.th <p>We model the optimization of cash logistic between a cash center of a commercial bank and the Bank of Thailand in Phuket using a mixed integer linear programming (MILP) problem, which is solved via a Microsoft Excel add-in OpenSolver.&nbsp; Conditions on the cash transport decisions are written as MILP constraints.&nbsp; Historical data of cash withdrawals and deposits between cash centers and the Bank of Thailand are input to a forecasting model (the AAA version of the Exponential Smoothing), which allows for seasonality.&nbsp; Our results show that our forecasts have lower MAPE than the forecasts currently used by the cash center. &nbsp;Using a MILP model combined with a new forecasted value, the cost of cash transportation (transportation cost, sorting cost, cost of fund) reduces by 24 Percent&nbsp;decrease in April 2019, and the sorting cost is not over the budget of 100,000 Baht.</p> 2021-06-15T09:57:21+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/244041 Designing of a Planning and Scheduling Tool for Earth Moving Operation 2021-06-15T10:01:55+07:00 Thitiya Theerarangsikul thitiya.thee@ku.th Worawut Wangwatcharakul fengwww@ku.ac.th <p>The main objective of this study was to generate a tool for planning earth moving operations logistics. Each project has a different distance, maximum work period, and soil volume, also, the company’s soil transportation main restraint was the limited resources which could only provide a total of 2,100 trips available per month for all the ongoing projects. If the total trips of earth moving each month exceeds the available resource limit of the company, the company will use a more cost consuming outsource service. The researcher applied the theory of linear programming to solve the problem of the operation planning. Company’s data for 16 projects over a 2-years working period were used to create a mathematical model and the most efficient planning strategy were assessed through excel solver. The model has shown that it is possible to reduce the total project duration by 2 months and the total cost of logistics by 1.44 percent or 3,319,517 baht.</p> 2021-06-15T10:01:55+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/244045 Parallel Machine Scheduling in the Testing Process of Integrated Circuit 2021-06-15T10:09:21+07:00 Thanawat Wongkrue thanawat.wongkr@ku.th Worawut Wangwatcharakul thanawat.wongkr@ku.th <p>This research studied the Parallel Machine Scheduling in the Testing Process of Integrated Circuit to minimize the number of machines in process and improved machine utilization. Two research methodology had been applied in this research. The first method is to create the production scheduling by Mathematical Model and using Solver program in Microsoft Excel to find out the optimal solution. The second one is created by Hybrid Heuristic, EDD+LPT rules. The solution obtained from the Mathematical Model and Hybrid Heuristic is compared the efficacy index with current scheduling in order to reduce total tardiness, total Earliness and Total cost system. From the result of 575 jobs, 4 weeks. The Mathematical Model method can provide the good solution with shorter amount of time than using Hybrid Heuristic. The efficacy index of total cost system is reduced around 13.43% and total tardiness/earliness are reduced 47.24%, 73% respectively.</p> 2021-06-15T10:09:21+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR https://ph02.tci-thaijo.org/index.php/TJOR/article/view/244060 OPERATING ROOM SCHEDULING UNDER LIMITED ICU BEDS 2021-06-15T13:21:24+07:00 Thitirat Ruansirilert ru.thitirat@gmail.com Sorawit Yaovasuwanchai yong2happy@gmail.com <p>Many patients in public hospitals around the world have to wait for a long time for their operations due to Operating Room (OR) and/or downstream resources scarceness. Moreover, increasing the number of ORs or related resources is difficult because of expensive OR costs and limited budget in hospitals. From our knowledge, a well OR schedule can raise number of patient throughputs and maximize the OR utilization so, we developed a binary programming model which aims to maximize the number of patients that can be included in the OR schedule under the limited number of ICU beds. Our OR schedule specifies which patient will be operated, OR, day, and period. Then, ten instance sets were generated from real data to test the model and solved by an optimization software. The results showed that, for small and medium instances, our model can guarantee the optimal solutions under 3-hours limited time, while the large instances cannot. However, there are small gap percentages for non-optimal solutions.</p> 2021-06-15T00:00:00+07:00 Copyright (c) 2021 Thai Journal of Operations Research : TJOR